NLLB translation mask or skip some tokens

[ekmekovski]

İs there a way to skip some words in translation,I jus want to get the input in the same way at my result (ofc since gramattical differences the order will change but I need to get the exact words in output) I tried to wrap some words with some placeholders $$$skipping words$$, I also tried to creat enumerated placeholders so that i can change them after translation (1_1_1, <1>,<2>,....). I also tried to give numbers. directly Like 11111111, 11111112, ... but None of them worked. Did someone try to do similar think?
Thnx in advance